I am working on a new project where I want to include an interactive tool for fitting multiple Gaussians to a dataset. This will only be one tool of several. I envision the tool being made up of a graph with some heavily customized cursors, an additional normal graph, a list of the cursors, a few buttons, and an indicator of the fitting results. I am weighing the options of how to implement this. One option, which I am more familiar with, would be a fully encapsulated actor in a subpanel-- the other would be a cluster of controls managed by a Q-Control. I have some experience with using actors and subpanels in this way but am wondering if a Q-Control (or alternatively something like JDP Graph Augments-- just mentioning for completeness) might make more sense. I would appreciate some feedback on why people well versed with both approaches might choose one over the other.
